Misunderstanding 1: "We can never invent an AI that is comparable to human intelligence."
In some specific areas, such as chess, go, stock market trading, and human-machine dialogue, we have invented computers that rival or surpass the human brain. Computers and the algorithms that drive them will only get better, and computer proficiency is just a matter of time.
New York University psychologist Gary Marcus once said that almost everyone who studies AI believes that machines will eventually replace humans. “The only difference between fanatics and skeptics is the judgment of time.†The futurists, like Ray Kuzwell, believe that the machine ruled humans would happen decades later, while others said It has been for centuries.
The machine intelligence originally envisioned may lack one or more of these attributes. In the end, we may create an extremely intelligent AI, but it cannot feel the world through subjective self-awareness. Shanahan said that intelligence and consciousness may be combined in one machine, but we must also recognize that intelligence and consciousness are two different concepts.
The AI ​​skeptics believe that the biological brain is unique in nature. This is an unsolvable scientific and technological problem. In fact, these are not convincing. Our brain is a biologically meaningful machine, but it is just a machine. They exist in the real world and follow basic physical laws. There is nothing in the human brain that is unknowable.
Myth 2: "Artificial intelligence will be conscious."
A common assumption about machine intelligence is that machine intelligence will be conscious in the future. In other words, the machine will think like a human. Critics, such as Paul Allen, one of Microsoft's founders, further believe that we have not yet reached the artificial general intelligence that machines can perform any task that requires intelligence like humans because we lack scientific theories about consciousness. As Murray Shanahan, a cognitive robotic researcher at Imperial College London, points out, we should avoid confusing the two concepts of cognition and consciousness.
“Exploration awareness is indeed a fascinating and important topic, but I think that consciousness is not necessary for human-level artificial intelligence. Or more precisely, we use the word 'consciousness' to refer to the psychological sum. Cognitive attributes, these are tied together in humans."
The machine intelligence originally envisioned may lack one or more of these attributes. In the end, we may create an extremely intelligent AI, but it cannot feel the world through subjective self-awareness. Shanahan said that intelligence and consciousness may be combined in one machine, but we must also recognize that intelligence and consciousness are two different concepts.
It can't be considered conscious just because a machine passed the Turing test. The advanced AI left us with the impression that they are conscious, but the advanced AI will not know more than a stone or a calculator.
Misunderstanding 3: "Super artificial intelligence will be very friendly."
The philosopher Kant believes that intelligence and morality are inseparable. Neuroscientist David Chalmers draws on Kant's famous ideas in his essay "Extraordinary Philosophical Analysis" and applies it to the promotion of artificial intelligence.
If this view is correct, we can think of the situation in which AI has a sharp rise in morality due to explosive growth in intelligence. It is conceivable that ASI systems become super-smart and have a sense of morality, so we can think of them as good.
However, this "advanced AI will be good, open-minded in nature" is untenable. As Armstrong pointed out, there are many smart war criminals. The relationship between intelligence and morality does not seem to exist in humans, so he suspects that this hypothesis is difficult to establish in other forms of intelligence.
"The unethical behavior of smart people can cause greater damage than their stupid compatriots," he said. "Intelligence just gives them the ability to get worse. It doesn't turn them into good people."
Misunderstanding 4: "Simple program patching can solve the control problem of AI."
Assuming that we create a greater AI than humans, we will face a serious problem: control problems. How to place and control ASI at home? How to ensure that ASI is friendly to humans? On these issues, futurists and AI theorists are completely at a loss. Recently, researchers at the Georgia Institute of Technology naively believe that AI can learn human values ​​and social customs by reading simple stories. This may be far from being solved by reading a few short stories.
“People have recommended a number of simple tips that might solve the entire AI control problem,†Armstrong said. Examples include writing a program that pleases humans for ASI, or simply using ASI as a human tool. Or, we can integrate positive concepts such as love or respect into his source code. In order to prevent ASI from having a single and extreme view of society, we can write programs that appreciate intellectual, cultural and social diversity.
But these methods are either too simple, which is like trying to define a simple superficial definition of complex human likes and dislikes, or to rigidly compress complex human values ​​into a simple word or idea. For example, for the word "respect", it is extremely difficult to determine a consistent and generally accepted definition.
"It's not that these simple techniques are useless. There are a lot of these countermeasures that point the way to the research and help to solve this ultimate problem. But these are just ideas. The specific operation requires us to do more development research. Work," Armstrong said.
As described in "Mechanical Ji", it will be very difficult to control artificial intelligence that is smarter than humans.
Myth 5: "Artificial super intelligence is very smart and will not make mistakes."
AI researcher and Surfing Samurai robot developer Richard Loosemore believes that most of the hypothetical AI's destruction of the world is not logical. People always imagine that in this scenario, the AI ​​might say: "I know that destroying human civilization is just my design failure, but I have to do it." Loosemore pointed out that if AI is going to destroy humans like this, it will have a lot of life. Such logical contradictions will also destroy their own knowledge base, causing them to become stupid and not enough to harm us. He also asserted that those who thought that "AI would only do things according to the specified procedures" made the same mistakes as those who used to say that computers were not flexible.
Peter McIntyre and Stuart Armstrong, who work at the Future of Humanity InsTItute, have different opinions. They believe that artificial intelligence will be largely limited by programming. They believe that AI also makes mistakes, or they are not smart enough to understand what humans want from them.
"Definitively, Super Artificial Intelligence (ASI) is a carrier of intelligence that is smarter than the best human brain in all relevant fields." McIntyre told Gizmodo. "It will be very clear what we will do with them." McIntyre and Armstrong believe that AI will only do the task of programming. If it can become smart enough, it will understand the essence of the task and the rules. Different, will also be able to understand human intentions.
McIntyre compares the future of humanity to the plight of mice. Rats want to seek food and habitat in human habitats, and humans do not want to have mice in the house. The two conflict. McIntyre said, "As we understand the purpose of the mouse, the super-intelligent system can also know what humans want, but it is still indifferent to it."
Misunderstanding 6: "We will be destroyed by super artificial intelligence."
There is no evidence that we will be destroyed by AI or unable to control them. As AI researcher Eliezer Yudkowsky said: "AI does not hate or love, but human beings are made up of atoms. In addition to intelligence, humans still have feelings."
Nick Bostrom, a philosopher at Oxford University, mentioned in his book Super Intelligence: Paths, Dangers and Strategies that once a true super artificial intelligence is completed, it will be more dangerous than any invention of humans in the past. Some well-known scholars such as Elon Musk, Bill Gates, and Stephen Hawking (who warned that AI may be the worst mistake in human history) also warned.
But nothing can be determined. No one knows what action AI will take and how it threatens humanity. As Musk pointed out, artificial intelligence can be used to control, regulate, and monitor other artificial intelligence. It may be possible to instill human values ​​and make them friendly to humans without reservation.
Misunderstanding 7: "We should not be afraid of AI."
In January, Facebook founder Mark Zuckerberg said that we should not be afraid of AI, AI will bring incalculable benefits to the world. Zuckerberg only said half right. We are always ready to harvest huge benefits from AI – from autonomous vehicles to the development of new drugs, but no one can guarantee that all the effects of AI are benign.
A highly intelligent system may understand everything about a particular task, such as solving annoying financial problems or hacking into an opponent's system. But beyond these special areas, highly intelligent systems become extremely ignorant. Google's DeepMind system is proficient in Go, but after Go, it has no other reason to explore the rationality and ability.
Using the Flame virus for cyber espionage in Middle Eastern countries
Misunderstanding 8: "The risks posed by AI and robots are the same."
This is a particularly common misconception that media and Hollywood movies such as the Terminator series, which do not identify the release of news, make this mistake even more popular.
If a super artificial intelligence such as Skynet wants to destroy humans, it will not use a robot with a machine gun. In fact, there are more efficient methods, such as spreading the plague, causing disasters such as grey goo (small robots with self-replicating ability), and even destroying the atmosphere. Artificial intelligence is potentially dangerous, not because of the future of the robot it indicates, but its self-positioning in the world.
Misunderstanding 9: "The AI ​​in the future is the same as the AI ​​in science fiction."
Indeed, science fiction works are used by writers and futurists to predict a better future, but the picture that ASI brings is not the same thing. More importantly, AIs that differ from human nature do not allow us to know and predict their exact nature and form.
Science fiction works that take into account that we humans are weak, and most AIs are set to resemble humans. “There are some views that even in humans, there are many types of mental models, and you and your neighbors are different, but the mental distinction between people is different from other possible mental forms. It’s trivial,†McIntyre said.
Most science fiction aims to tell a fascinating story, but not scientifically accurate. Therefore, conflicts in science fiction often exist between entities with similar strengths. Armstrong said, "The AI ​​without consciousness, joy or hatred destroys all human beings without any resistance and achieves a boring goal. It can be boring to imagine such a movie plot."
Misunderstanding 10: "AI will take away all our work."
Artificial intelligence automates most of our work and it can destroy the potential of humans. These are two completely irrelevant things. According to MarTIn Ford, author of The Rise of Robots, these two things are often confused. Thinking about the meaning of AI in the distant future is good, but it only makes sense if it does not affect the problems we face in the coming decades. The most important of these is large-scale automation.
There is no doubt that artificial intelligence will replace many existing occupations, from workshop workers to advanced white-collar workers. Some experts predict that half of the work in the United States will be replaced by automation in the near future. But that doesn't mean we don't have the ability to cope with the situation. The most ideal situation is to achieve the goal of praise for all human beings and to reach the realm of utopia. No matter whether it is a mental worker or a manual worker, everyone is liberated from labor.
Robots will replace millions of jobs in the coming decades. We must emphasize: "The AI ​​will kill a lot of work in the next few decades, but this is a good thing," Miller said. For example, autonomous vehicles will replace truck drivers, which will save on transportation costs and make the price of goods cheaper. "If you make a living by being a truck driver, you are unemployed, but everyone else earns because their wages buy more," Miller said. “The money saved by these people is used for other goods and services, which can create new jobs for humans.â€
Among all the possibilities, artificial intelligence will create new ways of creating wealth while freeing humans to do other things. Advances in artificial intelligence will lead to advances in other areas, especially in manufacturing. In the future, it will be easier to meet our basic needs, not the other way around.
?
Interchangeable Power Adapter: Simplifying Your Power Needs
As a leading manufacturer of ac power adapter, we understand the importance of providing efficient and versatile power solutions for our customers. In today's fast-paced world, having an interchangeable plug power adapter is essential.
1. Travel-Friendly:
For frequent travelers, carrying multiple chargers for different countries can be inconvenient and space-consuming. Multi plug power adapter is the perfect solution for those on the go.With EU US UK AU 4 plug detachable, you can travel to USA Europe UK Australia and charge more convenient.
2. Safety and Efficiency:
Safety is a top priority when it comes to switching power supply. Our changeable power adapter is built with advanced safety features, including over-voltage protection, short-circuit protection, and over-current protection, to safeguard both the adapter and the connected devices. With safety and efficiency in mind, our interchangeable power adapters offer peace of mind to users.
In conclusion, interchangeable power adapters offer a convenient and versatile charging solution for users. Their ability to work in various countries, travel-friendly design,emphasis on safety and efficiency make them an ideal choice for individuals seeking a simplified charging experience. As a trusted switching power supply manufacturer, we are committed to providing high-quality interchangeable ac adapter that meet the evolving needs of our customers. Choose our ac to dc power adapter and enjoy the convenience they bring to your charging routine.
Interchangeable Power Adapter, multi plug power adapter, interchangeable plug power adapter, detachable power adapter
Guangdong Mingxin Power Technologies Co.,Ltd. , https://www.mxpowersupply.com